Voice authentication leader Pindrop expands its AI-powered deepfake detection technology into healthcare, addressing growing synthetic voice fraud threats targeting patient data and medical systems.
Voice AI leader ElevenLabs will leverage Google Cloud services powered by Nvidia chips, expanding its synthetic audio infrastructure for next-generation voice cloning and generation.
New research introduces Constricting Barrier Functions for mathematically guaranteed safe outputs from generative AI models, offering formal safety proofs for controlled content generation.
New research introduces MINAR framework for understanding how neural networks learn to execute algorithms, advancing interpretability methods critical for AI safety and verification.
Researchers propose combining self-consistency sampling with conformal calibration to certify AI agent reliability without requiring access to internal model weights or architecture details.
New research introduces Tool-R0, a framework enabling LLM agents to autonomously learn tool usage through self-evolution, eliminating the need for curated training datasets while achieving state-of-the-art performance.
New framework converts opaque neural network decisions into interpretable mathematical expressions, enabling better model verification and understanding of AI behavior.
OpenAI has reportedly hired Apple's former head of AI models from Meta, intensifying the high-stakes talent competition among tech giants racing to develop frontier AI systems.
Anthropic acquires computer-use AI startup Vercept following Meta's poaching of a co-founder, signaling intensifying competition in the AI agent space.
New research reveals a surprising detection gap: while machines excel at spotting deepfake images, humans consistently outperform AI systems when identifying synthetic videos.
How modern AI models process massive context without quadratic memory explosion: Sparse Attention, Linear Attention, State Space Models, and Memory-Augmented transformers explained.
New research introduces SO-LoRA, combining sparse and orthogonal low-rank adaptation to enable efficient multi-task LLM fine-tuning over wireless networks with reduced interference.